* remove auto generated main method comment
* remove extra blank lines
* introduce CPU fallback if no GPU devices are available. include labels.txt in build copy
* rename getLMDFromAdapterIndex
* change d3d feature level to 11
* don't display the basic render driver to the user
* all c++ samples with print results have the correct output now
* c# samples are updated
* fix top3 bug for javascript sample
* update readmes to have correct output of top 3 probabilities
* replace locally declared struct with std::pair
* intermediate step in readme, need to check image rendering on github
* finish up debug_readme.md. restructure directory of customize_model
* debug_readme formatting
* add license file link
* rename custom operator sample directory. update base readme to include sample
* throw in a link in base readme for adapterselection which was forgotten in original adapter selection pull request
* update readme on future plan of debug operator
* fix conversion from relative to absolute path
* add description of text output format
* remove line about future plans for debug operator
* intermediate step in readme, need to check image rendering on github
* finish up debug_readme.md. restructure directory of customize_model
* debug_readme formatting
* add license file link
* rename custom operator sample directory. update base readme to include sample
* throw in a link in base readme for adapterselection which was forgotten in original adapter selection pull request
* update readme on future plan of debug operator
* fix conversion from relative to absolute path
* add description of text output format
* added .NET Core squeezenet example
* edit readme to acknowledge async limitations of .NET core 2
* refactor out async helper
* update readme
* remove c++ instructions copied from different README
* update Readme. remove liberal new lines
* parametrize assembly reference for Windows.winmd and update README accordingly.
* custom op sample
* finish customop cpu sample changes
* clearup straggling HMODULE variable
* merge in debug operator code. not running yet
* set working directory in vs project and copy debug model
* correct file name of labels file and lookup. Upgrade version of noisy relu custom operator.
* add comment
* set all baseline version to 7, and pass in different versions for each operator as not overwrite the operate schema set
* always return int in main method
* include python script example to modify onnx models programmatically
* update python script to be two scripts, one that inserts a debug operator into the given output and another that will insert a debug operator for every output.
* implement pull request comments from branch open on mscodehub.windowsai.
* switch tabs to 4 spaces
* include in the readme instructions for reusing the debug operator code
* update readme to be more verbose
* include example models output from python scripts
* resize headers in readme
* use argparse in python scripts. add comments
* remove unnecessary newlines remove comments
* remove blank lines in python scripts
* use path.join for separator in python script
* adding DeviceTarget solution to this fork. For a successful run you will need to move the Labels.txt file into Debug
* user input consumed for adapter selection
* rename DeviceTarget to AdapterSelection and restructure directory
* add readme
* remove unnecessary braces, arg parsing
* select input and output features dynamically.
* initialize session with correct device. target results using non hardcoded output name
* make build consistent between release/debug and 86/64. Handle user input that is not an number correctly
Sample which allows the user to select from an enumerated list of device adapters, and then constructs a LearningModelDevice from a D3D12 command queue .
* added Emoji8 source code to Samples folder
* simplified session object creation in the IntelligenceService
* added default values and modified countdown if statement
* added aka.ms links and logo to readme
* Update readme.md
LoadModel changed to LoadModelAsync to represent good practice in naming async models,
Removed void and replaced with async Task when loading the model
Ori Levari
Nick Geisler
smk2007
Ori Levari
Kumar Rajeev
Killian McCoy
Kumar Rajeev
Eliot Cowley
Matt Wojciakowski
Xiang Zhang
Young Kim
Young Kim
Serena Zheng
Nick Geisler
Paul McDaniel
Paul McDaniel
Rosane Vallim
Louis-Philippe Bourret
Shu-Wei Hsu
serenaz
Wilson Lee (CLOUD AI)
Eric Snyder
Eric Snyder
Louis-Philippe Bourret